Vm-dev March 2008

vm-dev@lists.squeakfoundation.org

17 participants
19 discussions

Profiling and delay accuracy
by Andreas Raab 06 Aug '11

06 Aug '11

Hi Folks - For a long time now I had this feeling that MessageTally reports pretty bogus results in many situations without being able to precisely track down why this would be the case. Today I was looking at primitiveResponse() and, much to my surprise, noticed that the checkForInterrupts() that used to be there for years is gone (see Squeak3.6 for example which still has it). The sole reason for this check was to get both high delay accuracy and good profiling behavior. Consider this example: fc := FormCanvas on: Display. MessageTally spyOn:[ 1 to: 100 do:[:i| fc fillColor: Color white. fc printString. ]. ]. Using 3.6 this will be measured correctly as: **Tree** 87.5% {243ms} FormCanvas>>fillColor: |87.5% {243ms} GrafPort>>fillRect:offset: | 87.5% {243ms} GrafPort>>copyBits 12.5% {35ms} FormCanvas(Object)>>printString 12.5% {35ms} FormCanvas(Object)>>printStringLimitedTo: whereas the 3.10 reports it as **Tree** 99.6% {670ms} FormCanvas(Object)>>printString 94.9% {639ms} primitives 4.6% {31ms} FormCanvas(Object)>>printStringLimitedTo: I think you can see the problem and just imagine what this does to any cumulative results. I also wonder what that change does to platforms that don't have high-resolution asynchronous timers. This used to be the one place where we would catch timer changes at high resolution and unless the platform forces high resolution timer interrupts asynchronously (by resetting the interrupt counter) we must be loosing a lot in terms of delay accuracy. (it just so happens that the Windows VM *does* do asynchronous timer interrupts so I never noticed this but I'm very curious about the difference on Linux which typically still has pretty bad timer granularity). Any opinions? My feeling is that the timer check should go back in for all of the above reasons. At the very least a quickCheckForInterrupt() is required but I actually think putting the full check back in is worth it. Cheers, - Andreas

3 3

Joys of Squeak-dev image, VM Compiling on MacIntel and Spoon remote OtherBrowsing
by Ken G. Brown 16 Nov '10

16 Nov '10

MacPro OS X 10.4.9 Xcode 2.4.1 Squeak-dev-114-2, VMMaker 2.8b6.1, Squeak 3.8.15beta6U.app, and version 1657 of svn MacOS VM sources, and Spoon v2a12. Downloaded http://www.squeakvm.org/cgi-bin/viewcvs.cgi/ trunk version 1657 for MacOS Ran squeak-dev-beta-114-2.image on Squeak 3.8.15beta6U.app Loaded VMMaker 3.8b6.1 from dev image Universes. Generated vm sources by following instructions in <http://www.squeakvm.org/cgi-bin/viewcvs.cgi/trunk/platforms/Mac%20OS/vm/Doc…> Fought with Xcode wanting to use MacOSX10.3.9.sdk for ppc which was giving an error which was known and documented in the Readme. Finally figured out what Xcode wanted so built universal app just using 10.4.9 sdk. Couple issues: Exupery references remain in sqnamedPrims.h and sqPlatformSpecific.h from svn version 1657, I just commented them out. Notes said: Info-NPSqueak__Upgraded_.plist was to be found in the Developer directory. I could not find it so scammed a info.plist out of John's Squeak 3.8.15beta6U.app vm. Not sure what impact this has. And much, much diddling about figuring out various things... Finally got a build. It's a lot smaller (3.9MB) than John McIntosh's Squeak 3.8.15beta6U.app (8.1MB), and it looks like it is a PowerPC one but at least it runs. I'll look into the Xcode build process some more. Then: Replaced interp.c with Craig's interp.c for Spoon, rebuilt and got a VM with Spoon client stuff. Downloaded Spoon release 2a12. Moved a copy of Flow plugin from within Craig's naiad.app package into a plugins subdirectory where the new spoon client compatible MacIntel universal VM was run from. Followed Craig's Spoon instructions for loading niaid.st and niaid.pr into the 3.10 image (Squeak-dev-114-2) running on the new VM. Did the run spoon.app thing from Spoon.dmg (version 2a12), needs to be done twice if Safari is not running. Got Spoon processor running. Followed instructions again in the Spoon client and lo! Got OtherBrowser browsing the Spoon memory running on processor. Yay Craig! Who do i get to sign my certificate? :) Ken

4 7

How to use Croquet VMMaker
by Matthew Fulmer 29 Apr '09

29 Apr '09

Hi. In anticipation of working on HydraVM, I am trying to test as close a configuration as possible to HydraVM without being hydra. That is, load the nearest ancestor VMMaker and C sources to hydra, and try to make a working VM. This should help ensure that there are no "normal" bugs once I switch to Hydra, only hydra bugs. So, I did that: Here is my setup: - Fedora core 7 on x86 processor - Release 3.9 image/VM from squeak.org - Load FFI from SqueakMap - Load VMMaker-ar.70 from http://jabberwocky.croquetproject.org:8889/Homebase/ - Checkout http://squeakvm.org/svn/squeak/trunk revision 1796 (The nearest ancestor of HydraVM) I pointed VMMaker at the platforms directory in the checkout, and told it to build source in platforms/unix/src. I selected all plugins to be external, but moved FilePlugin and SocketPlugin to be internal. Then I did mkdir unixbuild cd unixbuild ../platforms/unix/config/configure make make fails linking gnu-interp.o: /bin/sh /home/matthewf/squeak/vm/hydravm-trunk-base/unixbuild/libtool --mode=link gcc -g -O2 -fomit-frame-pointer -DLSB_FIRST=1 -export-dynamic -R/usr/local/lib -o squeak vm/vm.a FilePlugin/FilePlugin.a SocketPlugin/SocketPlugin.a disabledPlugins.o version.o -lutil -ldl -lm -lnsl vm/vm.a gcc -g -O2 -fomit-frame-pointer -DLSB_FIRST=1 -o squeak disabledPlugins.o version.o -Wl,--export-dynamic vm/vm.a FilePlugin/FilePlugin.a SocketPlugin/SocketPlugin.a -lutil -ldl -lm -lnsl vm/vm.a -Wl,--rpath -Wl,/usr/local/lib vm/vm.a(gnu-interp.o): In function `primitiveGetWindowLabel': /home/matthewf/squeak/vm/hydravm-trunk-base/unixbuild/vm/gnu-interp.c:15141: undefined reference to `ioGetWindowLabel' vm/vm.a(gnu-interp.o): In function `primitiveSetWindowLabel': /home/matthewf/squeak/vm/hydravm-trunk-base/unixbuild/vm/gnu-interp.c:18770: undefined reference to `ioSetWindowLabelOfSize' collect2: ld returned 1 exit status Building this VM worked when using the latest VMMaker from SqueakMap (VMMaker-tpr.58), everything else the same. Is there something else I need to do in order to build using the Croquet VMMaker? I want to use it specifically, since the Hydra VMMaker is derived from it. The dev notes on Hydra are at http://squeakvm.org/~sig/hydravm/devnotes.html -- Matthew Fulmer -- http://mtfulmer.wordpress.com/ Help improve Squeak Documentation: http://wiki.squeak.org/squeak/808

5 9

Interpreter>>signed32BitValueOf: and signed64BitValueOf: broken
by Andreas Raab 26 Nov '08

26 Nov '08

Hi - I just noticed hat Interpreter>>signed32BitValueOf: and signed64BitValueOf: are broken for edge cases. The following example will illustrate the problem: array := IntegerArray new: 1. array at: 1 put: 16rFFFFFFFF. "should fail but doesn't" array at: 1. "answers -1 incorrectly" array := IntegerArray new: 1. array at: 1 put: -16rFFFFFFFF. "should fail but doesn't" array at: 1. "answers 1 incorrectly" The problem is that both signed32BitValueOf: as well as signed64BitValueOf: do not test whether the high bit of the magnitude is set (which it mustn't to fit into a signed integer). The fix is trivial in both cases - basically all that's needed at the end of both functions is this: "Filter out values out of range for the signed interpretation such as 16rFFFFFFFF (positive w/ bit 32 set) and -16rFFFFFFFF (negative w/ bit 32 set). Since the sign is implicit in the class we require that the high bit of the magnitude is not set which is a simple test here" value < 0 ifTrue:[^self primitiveFail]. negative ifTrue:[^0 - value] ifFalse:[^value] Cheers, - Andreas

4 12

FFI 64-bit updates - progress report
by David T. Lewis 19 Apr '08

19 Apr '08

I now have FFI working on a 64-bit AMD Linux system for both 32-bit and 64-bit images. All unit tests pass, including some new tests that I added to provide more coverage of the FFI primitives. The X11Display examples also work. This is using the "cdecl" interface in platforms/unix/plugins/SqueakFFIPrims/lib-anyffi.c. The next step will be to move the changes to a 32 bit Linux system and do the updates to platforms/unix/plugins/SqueakFFIPrims/x86-sysv.c. This is work in progress, but I'm attaching a snapshot of the work that I have done so far. There are changes to the VM, the FFI plugin, and to the FFI support classes in the image (note: after applying the changes, classes with cdecl pragmas must be recompiled). More to follow, Dave

1 2

Re: [squeak-dev] new display driver for Syllable
by David T. Lewis 01 Apr '08

01 Apr '08

He Kelly, (CC to the vm-dev list) I don't know the answer to your question, but there is a separate mailing list devoted to Squeak VM development that may be of some help. http://lists.squeakfoundation.org/mailman/listinfo/vm-dev - Dave On Sun, Mar 23, 2008 at 10:13:12AM -0600, Kelly Wilson wrote: > Hello everyone, > > > > I recently ported squeak to Syllable OS and it runs with vm-display-null and > vm-sound-null just fine. Now I would like to develop the video driver for > Syllable.unfortunately Syllable's windowing system is written in C++ and > interfacing to it could be difficult (they may have a C interface set > up.I'll check). As a backup I was wondering if I can just use SDL by > modifying the vm-display-custom? > > > > I tried to modify the vm-display-custom code and it compiles fine with some > SDL functionality but I have one weird problem. If I add any external call > into the sqUnixCustomWindow.c file then compile/link/install it, the squeak > module loader can't find vm-display-custom anymore. I can modify the file by > adding some structures and functions local to the file but, like I said, if > I call outside this file (and then include -lSDL or whatever when linking) > then the loader just can't even see vm-display-custom. > > > > I can see display_custom and display_custom_itf with the 'nm' command and > the format of the shared lib looks fine??? Does the loader look for > something really specific about the libs size or something strange like > that? I assume it dlopens() it. > > > > Thanks, > > K.Wilson > > > > P.S. I am sure that I have all the libs and things in the correct places > because I can just recompile by commenting out my single external function > call, recompile, reinstall and things work fine. > > > >

3 3

Re: How difficult to make a .NET VM?
by Aran Lunzer 27 Mar '08

27 Mar '08

Hi Andreas > One thing to experiment with the -browserWindow: > command line argument. Originally, it was designed to "get Squeak > into a browser" but of course, as long as your .NET form has a > HWND handle you can get Squeak inside it by passing the handle > along with -browserWindow:. Well, b*gger me. If I'd known this (uhh, perhaps if I'd asked this) a year ago... Anyway - thank you. It seems to work: Squeak will happily run inside a .NET form. Next challenge: I want to tell that .NET form (call it the substrate) to be the owner of other forms created within the application - because that way they'll be guaranteed (in theory) to stay in front of the Squeak window. But to be able to talk to the substrate at all, the Squeak VM must be in the same process as it. How do you launch Squeak in the same process as the form? Well... remarkably enough, the following seems to work: 1. start Squeak 2. in Squeak, crank up the .NET bridge 3. use the bridge to create a free-standing, Squeak-desktop-sized .NET form 4. tell the VM to use that form as its browserWindow ...but I'm not yet confident that this approach would fly for the case of a plugin within a real browser. I'll have a go. And meanwhile I have another question: Is there a standard way of sending messages between the VM and Squeak code? For the time being I've hacked together a simple shared-data link using the GWL_USERDATA value of the Squeak window, as well as a one-way signal from VM to Squeak using some custom UI-event types. My next thought for talking to the VM was to try posting custom messages to the Squeak window's queue. Any other mechanisms? Thanks again Aran -- Aran Lunzer Hokkaido University

1 0

Tim's New Compiled Method format
by Matthew Fulmer 27 Mar '08

27 Mar '08

While talking to Craig about spoon, I learned that Tim made a new Compiled Method format using normal objects, and (I think) a VM that ran it. I tried searching for it and came up with: New Compiled Method Format and 3.0 Image http://wiki.squeak.org/squeak/750 VI4 http://wiki.squeak.org/squeak/2119 Version 4 http://wiki.squeak.org/squeak/3716 I have been discussing what would go into Squeak 4 since we now need to create such a system, due to the urgency of the license situation, and I am now leaning toward Spoon rather than KernelImage as a basis for Squeak4, after learning the magnitude of the relicense effort on Monday. So, where could I find this code? Does it work? Is there just a CompiledMethod format change, or is there a whole other VM project I never heard of? I can't really tell, but now would be the time to use it if it works. -- Matthew Fulmer -- http://mtfulmer.wordpress.com/

4 6

How difficult to make a .NET VM?
by Aran Lunzer 27 Mar '08

27 Mar '08

Hi [Apologies if this has been addressed before or elsewhere; my Google attempts didn't turn up anything that seemed relevant.] Has anyone tried packaging the Squeak VM as a .NET class? (...if that's a sensible way to ask the question) For the past couple of years I've been using and hacking the Squeak/.NET bridge, created by John Pierce and Ben Schroeder, mainly so I can create and control IE-like Web browser views within the Squeak desktop (see, for example, http://km.meme.hokudai.ac.jp/people/aran/papers/UIST06-LunzerHornbaek.pdf). While this has been successful to the degree needed to be a platform for HCI research, the setup is still - frankly - an untidy and fragile hack. The main challenge has been how to combine the .NET views with the Squeak desktop. Originally we made them children of the Squeak window, but this gave us a nasty, intermittent thread-based deadlock. We worked around this by keeping the .NET views independent but making them hover in front of the Squeak window in the Windows Z order; some tweaks to the handling of focus-change messages in the Squeak VM allow us to ensure - most of the time - that the combination shares the desktop nicely with other Windows applications. However, as I say, it's pretty fragile. So we're wondering whether there's a clean way to make the .NET views be children of the Squeak window, only without running into the COM vs. .NET threading issue that hit us before. In particular, unless we can crack this issue it seems unlikely that we can get the Squeak+.NET combination to behave acceptably when embedded within a Web browser. I suspect that if we could package the Squeak desktop itself as a .NET Form we could then make the other views be proper children of that desktop, and the window management problems would all go away. I also imagine that we could then go the extra step of packaging this .NET-based VM as a browser plugin. If it all works, the result would be a pretty useful (though Windows-specific) Squeak+.NET environment. So: can anyone shed some light on how complex this is likely to be? Has anyone already tried it? ...or would we be better off just looking for another way to get Web pages rendered? A Squeak-native approach such as Scamper seems doomed to lag the state of the art by years; has anyone looked into packaging a library such as Gecko for use within Squeak? Thanks for any thoughts - Aran -- Aran Lunzer Hokkaido University, Japan http://km.meme.hokudai.ac.jp/people/aran/

2 1

callback patch
by Gwenaël Casaccio 26 Mar '08

26 Mar '08

Hi, I try to compile the callback patch but there is a problem when I launch the vm there is a nice black screen of death :) and the vm is frozen. Do you have any idea of the problem ? Cheers Gwenael Casaccio

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev March 2008